When it comes to the utilization of commercially available advanced electronics packages in military applications, the manufacturing technology challenge facing the staff at the EMPF has always been and continues to be the identification and mitigation of high risk areas affecting the reliability of commercial off the shelf (COTS) electronic components in harsh environments. The EMPF, partnering with Raytheon Electronic Systems and Rockwell Collins, successfully completed a Navy ManTech project in 2002 titled Electronic Miniaturization for Missile Applications (EMMA), which determined, evaluated, and mitigated the risks associated with integrating miniaturized electronics packaging technologies into existing weapons systems. As a requirement of the EMMA project, advanced electronics packages (figure 2-1) were investigated and failure analysis data was compiled and summarized in the Technical Applications Guidelines (TAG) Handbook.

The EMMA project proved the following guidelines for improving solder joint reliability:

• Match the Coefficient of Thermal Expansion (CTE) of the component and board material.
• Apply an underfill to the Chip Scale Package (CSP) and Flip Chip (FC) packages.
• Use smaller and lighter components.
• Do not use ceramic components in environments with large temperature extremes.

Following these recommendations will help to reduce but not completely eliminate solder joint failures. Identifying the root cause of failure will be dependent upon a variety of factors. Several variables to consider when performing failure analysis on advanced electronic packages are solder joint cracking, package assembly variables, and solder voids .

Solder Joint Cracking
Solder joint cracking is an indication that the solder joint cannot withstand the stress of its environment. A component's size, weight, and location on the board are all factors affecting the reliability of the solder joint. For example, the CTE differences between the component and the board can contribute to the rate that a solder joint cracks (Table 2-1).
Cracks can appear at the component/ solder interface or the board/ solder interface (Figure 2-2).

Package Assembly Variables
Failure analysis performed during the EMMA project discovered that the individual component construction of the electronic package assembly selected contributed greatly to the solder joint performance. As one team member pointed out, not all Ball Grid Array (BGA) assemblies are equal. One example, the flexible chip scale package, had a component construction that impacted the solder joint thermal cycle fatigue life. The component pad used a solder mask defined pad construction that significantly reduced the cross-sectional area of the solder joint geometry. This reduced area caused a stress concentration at the solder ball-component interface, where cracks within the solder joint were propagated.

A second example, a dimple ceramic BGA (CBGA) in which the solder balls were recessed into the package, had a component construction that yielded poor reliability due to the CTE differences between the ceramic package and the board. As with the plastic BGA with solder masked defined pad construction, the ceramic device tested had cracks at the solder ball-component interface.

A third example, a "sagged" solder ball component interface with a taped BGA, showed that the component construction contributed to solder joint failures at the component pad-solder joint interface and were attributed to "sagging" located on the pads on the outer rows of the component and not directly beneath the component die. The majority of the solder joint failures were located at the component pad-solder joint interface.

Solder Voids
Solder voids have always been a reliability problem in electronics manufacturing. According to IPC-A-610, Acceptability of Electronics Assemblies, up to 25% voiding is permitted in a solder ball at the board-to-ball interface. Typically solder voids indicate that there is a problem with the thermal profile of the soldering process. However, failure analysis performed during the EMMA project discovered that package construction may also contribute to solder voids. The volatilization of an organic material (PWB fabrication chemistry, solder mask residue) captured in the via can contribute to solder void creation. Volatilization of solder paste flux trapped by the interaction of surface tension forces with the via geometry can also generate solder voids (Figure 2-3).

It is possible that a solder void, acting as a stress relief, can prevent a crack from propagating through the solder joint. However, a void that is too large can produce an unreliable solder joint.

Conclusion
As the EMMA project demonstrated through rigorous analysis, it is feasible to incorporate advanced electronic packages into existing military systems requiring high reliability in harsh environments. By employing testing and analytical methodologies such as failure analysis as part of the manufacturing evaluation and control process, the root cause of solder joint failures can be identified by considering the construction of the component, the properties of the materials,and the soldering process.

Acknowledgements
The author would like to acknowledge the contributions from J. Silva, Raytheon, and D. Hillman, Rockwell Collins for their activities on the EMMA Project.